The ?UCSC-Buck Genome Data Analysis Center for the Genomic Data Analysis Network? will develop state-of-the-art methods for integrating various types of data to discover the genetic pathways, the microenvironment, the originating cells, and the oncogenic processes driving the initiation and progression of tumors. The long term goals of the project are to identify highly accurate models detailing the faulty genetic circuitry at work in each subclone of a patient?s tumor, as well as any ?normal? cells acting as accomplices by supporting the cancer microenvironment. The ultimate objective is to encode computer algorithms that can search a patient?s individual pathway diagram for the best combination of interventions to eliminate every tumor cell, while preserving the health of every normal cell, in their body. Integrative pathway analysis methods will be developed to reveal signatures of tumor subtypes from Pan-Cancer and external datasets. New technologies will be established for uncovering network models tailored to individual patients. The tools will be deployed as part of an active collaboration to support the specific projects of the Genome Data Analysis Network. Novel probabilistic graphical models will be used to infer disrupted signaling. Cellular signatures will be collected from the analysis of normal cells, cancer cell line models, and Pan-Cancer investigations. Novel machine-learning methods, guided by pathway mechanisms, will be established to identify cell state signatures in heterogeneous patient samples. This work will reveal rare mutations driving metastatic transformation that are currently of unknown significance. New clues about the genetic circuitry promoting response and resistance to treatment will be established. Finally, cross-tumor connections that relate tumors of one type to a different type will suggest new avenues for treatment.